Written by: Vaidik Mandloi , thetokendispatch

Compiled by: Plain Language Blockchain

In January 2026, an anonymous trader placed a series of bets on the cryptocurrency exchange Polymarket, wagering that Venezuelan President Nicolás Maduro would be captured. The total stake was approximately $34,000. A few days later, U.S. special forces carried out the capture operation, and the trader cashed out over $400,000 in bets. The Secretary of State later confirmed that the operation was too sensitive to require congressional notification. Consider this: the U.S. Congress, responsible for authorizing military operations, was completely unaware. The American public was also completely oblivious. Yet, someone sitting behind the screen of a cryptocurrency betting platform possessed ample information and wagered real money. And his prediction came true.

This has become a common saying in the prediction market industry today. As Shane Copeland, CEO of Polymarket, put it, it's known as the "truth machine." The argument is that because traders have a vested interest, their collective betting reflects the future direction of the world more accurately than any opinion poll, expert, or commentator (who bear no consequences even if their predictions are wrong). In short, Polymarket's odds are arguably the closest thing to the truth you can find.

This argument seems to be working. Prediction markets are no longer a niche corner of the internet frequented only by a few gamblers seeking thrills. A recent dataset analysis of 364 TikTok videos mentioning prediction markets found that 68% of them were unrelated to trading. People weren't gambling, but rather citing odds from these platforms in political debates, much like they used to cite polls. Polymarket appeared in approximately 70% of these videos. One 22-year-old TikTok user even used cryptocurrency betting odds to predict real-world trends in a political video, and a significant number of people agreed with him.

This is incredible. Two years ago, you simply couldn't believe this was happening. But one question no one seriously considered was: are these probabilities really worth trusting so much?

So I want to ask: How accurate are these markets really? What happens when odds start to influence the events they're supposed to predict? What will the future look like when the whole world treats betting odds as gospel?

How to score the prediction market?

Before analyzing the data, we first need to understand how to measure the effectiveness of a prediction market. Because most people never consider this question, and without it, all the hype surrounding Polymarket and Kalshi is just marketing gimmicks.

There's a scoring method called the Brier score. Meteorologist Glenn Brier proposed this method in 1950 to assess the quality of weather forecasts, as weather forecasters were (and still are) among the first professions that required serious attention to probabilistic predictions and made a living from them. The method is very simple. Suppose you predict a 90% chance of rain tomorrow, and it actually rains. That's a good prediction, and your Brier score is low. Now suppose you predict a 90% chance of rain tomorrow, but it's sunny and clear. That's a bad prediction, and your Brier score is high. A Brier score of 0 means your prediction was completely correct. 0.25 means your prediction was the same as flipping a coin. Any score above 0.25 means you'd be worse off than guessing randomly.

Why does this matter? Because when Polymarket tells you their market predicts Trump has a 60% chance of winning, and he ultimately wins, it sounds amazing in the headlines, but statistically speaking, one correct prediction means almost nothing. You need to evaluate the market's complete historical record on thousands of issues. That's where the Brill Score comes in. It's the only honest way to assess whether these markets are truly good at predicting election results.

A website called Brier.fyi does exactly that. They analyzed over 84,000 questions across Polymarket, Kalshi, Manifold, and Metaculus platforms. Brier's composite score on Polymarket is 0.047. That's quite a good score. Simply put, imagine a forecaster saying, "I'm 90% sure this will happen," and then predicting it with that accuracy every time.

Source: Brier.fyi

But here's the interesting part: the claim of a "truth machine" is beginning to crumble.

The score of 0.044 is the average of all products listed on Polymarket. In this case, the average plays a crucial role. If we break down the score according to what people actually bet on, we find that the ratings fluctuate greatly.

Science and Economics? Polymarket rates it an A. This market is based on CPI data, Federal Reserve interest rate decisions, and GDP data. These markets perform well because traders tend to have some financial knowledge, the data is verifiable, and there are institutional investors with genuine expertise in the relevant fields investing real money.

Politics? B+. Passable, mainly supported by the huge market of presidential elections, where billions of dollars flow in. Culture and technology? Worse. Much worse.

Then there's sports. The overall score for sports prediction markets across all platforms is only 0.325, or -D. Keep in mind, the probability of flipping a coin is 0.25. In general, the performance of sports prediction markets is worse than simply flipping a coin to predict each question. Think about it.

The category that attracts the most recreational bettors, and which Kalshi has been actively expanding (at one point, about 90% of Kalshi's trading volume was concentrated in sports betting), is a category that has been proven unreliable in the market.

Now, looking at the various markets, that's where the story gets even more interesting.

Polymarket once launched a market prediction about whether Bitcoin would reach $100,000 by January 2025. Bitcoin did indeed reach $100,000. While the market prediction was correct, it misjudged the probability for most of the time, remaining at a low confidence level for months before surging to near-certainty in the final stages. Its Brier score was 0.4909, or F. Remember, below 0.25 (equivalent to the probability of a coin toss), you might as well just guess randomly. And this market's Brier score was almost double that value.

The market reaction to Kamala Harris winning the 2024 Democratic presidential nomination was even more frenzied. Ironically, she ultimately won the nomination, and the market once again predicted the outcome correctly. However, the Brill Statistical Index (BBI) was a mere 0.9098. This number is utterly abysmal, and the emphasis cannot be overstated. The market has long been so confidently wrong in its predictions that even a correct final result couldn't salvage it. If you've been relying on this market for your decisions, you'll be misled throughout the entire campaign cycle, right up until the very last moment.

Now let's look at the other side of the story, because this isn't a simple one. The 2024 US presidential election was a real victory for prediction markets. While all major polls predicted a close race, Polymarket predicted Trump's approval rating to be around 60%. A Vanderbilt University study used a Bayesian time series model to compare Polymarket's predicted prices with national polls in seven swing states. The results showed that Polymarket was more accurate in all aspects.

So what does this tell us? Prediction markets excel at election forecasting. Especially in the largest and most liquid elections, where billions of dollars, tens of thousands of traders, and the general public converge on the same issue, predictions often outperform polls.

The problem is that election predictions likely account for only 2% of the trading volume listed on these platforms. Polymarket's 2024 presidential election market alone generated over $3.6 billion in trading volume, with 63,000 unique traders monthly. If you focus on congressional elections, state referendums, or any issue in the cultural, technological, or sporting sphere, bid-ask spreads can soar to 20% to 100% of the midpoint. Spreads in legislative and crisis-related markets can even approach 100%. Such large spreads mean the market is virtually clueless. It's simply two people making diametrically opposed guesses on the same issue, both with virtually no financial backing.

When fate begins to write its story

If the accuracy issue were confined to the predictive market ecosystem, it could be managed. Traders betting on bad markets would lose money, learn from their mistakes, and the system would improve over time. This is how all financial markets have operated for a century. The problem is that the odds are no longer a signal within the traders' inner circle, but rather information publicly available to everyone.

Over the past 18 months, major U.S. news outlets have incorporated market prediction data into their political reporting. The Wall Street Journal signed a formal agreement with Polymarket to include Polymarket's betting data in its news coverage. CNN began displaying Kalshi's odds on screen during its election night coverage. CNBC adopted the same approach. In December 2024, even Substack announced a direct partnership with Polymarket, enabling newsletter writers to embed real-time market data directly into their articles.

This is why the odds eventually appeared on TikTok. These figures traveled from Polymarket to The Wall Street Journal, then to cable news, then to Twitter, and finally to TikTok. By the time ordinary users saw these odds, they had been disseminated through enough authoritative channels that people perceived them as fact. What people accepted were figures that had already been "whitewashed" by mainstream media.

This is the root of the problem in market prediction: once odds are disseminated as news, they begin to influence what they are supposed to predict. This phenomenon has a specific name, which economists call endogeneity. Simply put, it means that the measurement results change the object being measured.

Let me give you a concrete example. Coinbase CEO Brian Armstrong is participating in an earnings call. He learns that Polymarket is executing a contract that stipulates whether he will mention certain phrases during the call. So, he modifies his original wording. The market should have been able to predict what he would say. However, his understanding of market dynamics changes his final statement.

Now, let's zoom out to the election level. In the 2024 US presidential election, a French trader known as "Theo" (or "the Polymarket Whale" in the media) bet on Trump to win and ultimately profited over $85 million. This wasn't just a lucky gambler. He commissioned a private poll independent of all public national polls, which showed Trump's actual performance far exceeded what the public polls indicated.

This is why his bets drove up prices across multiple platforms, subsequently being reported by the media I mentioned, including The Wall Street Journal, CNN, and political commentators on various platforms. Despite polls showing a close race, market predictions favored Trump. This single narrative shaped the opinions of millions in the final weeks. Commentators debated whether "smart money" possessed information unseen in the polls. Voters accepted this narrative, and ultimately, Trump won.

I'm not claiming Theo changed the election results. That's a far-fetched claim, and I can't verify it. What I really want to say is that anyone following this should be concerned: a trader with access to private polling data that others can't obtain is able to influence Polymarket's price movements, and then the Wall Street Journal and CNN repackage this price as the collective wisdom of the market and disseminate it. A good prediction market should aggregate a large amount of information from numerous participants into a clear signal. What happened in 2024 was that one person's exclusive polling data was whitewashed through Polymarket and re-disseminated as if it represented the consensus of thousands of traders.

If a trader can do this with $85 million, think about what people who truly have money and power will do.

In February 2026, Israeli authorities indicted at least two people, accusing them of using classified military intelligence to gamble on the Polymarket platform. They placed bets on contracts related to Israeli military operations before these operations were made public, with a potential profit of approximately $100,000. These individuals possessed security clearances and used information unavailable to the public for several days to wager on war. This was the first such case globally, demonstrating that prediction markets are fast enough, liquid enough, and anonymous enough to be used to monetize classified information in real time.

The Maduro deal mentioned at the beginning of this article follows the same pattern. Someone placed bets before the secret military operation took place and won over $400,000. Whoever it was, they either had inside information or were one of the luckiest bettors in the history of political gambling. We'll never know.

What happens when everyone believes the odds?

On the Polymarket platform, questions are answered within four days at the median time. The average answer time is 19 days, but this can be increased by a few long-trading markets. Most questions on the platform relate to the market movements of the week.

This shows that these markets aren't making any meaningful long-term predictions about the future. They're simply pricing in near-term events. Will Friday's vote pass? What will this person say tomorrow? Will Wednesday's numbers be higher or lower than expected? This information is certainly useful. But it's completely different from what people mean by predicting markets being "truth machines." The term "truth machine" usually implies that the market can tell you what the world will look like six months, a year, or even five years from now. But data shows that it simply can't. Not at all.

Prediction markets see 99% of trading volume concentrated in the final hours before an event is decided. Funds flood in during those last few hours when the outcome is almost certain. Moreover, these markets have even larger liquidity gaps. By the end of 2025, the combined weekly trading volume on the two major platforms, Polymarket and Kalshi, is projected to reach approximately $2.5 billion. Sounds staggering, right? But the daily clearing volume in the US options market alone is a staggering $760 billion.

Prediction markets account for only 0.05% of the total. The entire prediction market industry, regardless of the platform, contract, or category it covers, is negligible compared to the markets that institutions actually use for decision-making.

Here's the situation: prediction markets only work for a very specific type of problem: binary, high-profile, short-term events involving millions of dollars. But this is only a small fraction of the many problems these platforms actually offer. For the remaining 98% of problems, prices are unreliable, liquidity is virtually zero, and the results are more like Twitter polls than financial instruments.

They are building the default probability source for everything. Just as you open the Bloomberg terminal when you want to check stock prices, their vision is for you to open Polymarket when you want to check probabilities. Their strategy is that once enough media outlets, newsrooms, financial analysts, and government researchers rely on this data source, regardless of the accuracy of the data, the product will become irreplaceable.

I think this will work. And I think it should make everyone worry.

The question isn't whether predicting the market is useful. The answer is yes. For elections, major economic data releases, and a few high-profile events, they consistently outperform other alternatives. This is a fact, and crucially so. The question is, what happens when the entire information ecosystem begins to treat these market outputs as gospel, even for the 98% of questions that the market is completely unpredictable?

Economist Robin Hansen has been advocating for prediction markets for decades, describing them as a system that forces people to invest in their beliefs. In his model, the final price will be the best estimate of the current truth. However, this model assumes that the market is highly liquid, with diverse participants and is not easily manipulated. Our current markets, however, are dominated by a few "whales" concentrated in two areas (elections and sports), accounting for about 80% of total trading volume. The remaining 20% ​​of trading volume is spread across tens of thousands of contracts, where a few thousand dollars can cause double-digit price fluctuations.

These are all tools for generating attention. They work when the whole world is watching; they fail when no one is watching. The more people believe they are tools for creating truth, the more power those who can manipulate prices wield. And those who can manipulate prices are not a group of well-informed ordinary people, but a small group of wealthy traders who control private polls and, in at least two confirmed cases, have access to classified information.

The most dangerous aspect of market prediction isn't that it can be wrong, but that it consistently and precisely predicts key issues, thereby gaining a level of trust it didn't originally possess. This trust is gradually permeating the world's information processing mechanisms. The Wall Street Journal published the prediction data, CNN broadcast related content, and TikTok also shared the data. Ultimately, a trader with sufficient capital will determine the meaning of these numbers.

This is the reality of the truth machine. A system that generates numbers, which the world decides to call truth.